An image digital content rendering method, device and system

Abstract

This invention belongs to the field of image processing technology. It provides a method, apparatus, and system for rendering digital content in an image, comprising: acquiring all transparent objects in the rendering scene, classifying and labeling them and assigning unique IDs; dividing the scene into equal spatial blocks using an octree algorithm; equipping each spatial block with a unique set of classification labels; constructing a static transparent object depth value table; establishing a dynamic depth value prediction model by combining dynamic object motion parameters, animation keyframes, or interaction history; filtering effective spatial blocks through axis-aligned bounding boxes and view frustum intersection tests; generating an initial rendering sorting table by combining the maximum depth value of static objects and the depth stability weight of dynamic objects; filtering overlapping transparent object combinations based on oriented bounding boxes and the separating axis theorem; verifying spatial relationships using ray casting; and correcting the sorting order using a weighted scoring method.

Description

Technical Field

[0001] This invention belongs to the field of image processing technology, specifically a method, apparatus and system for rendering digital image content. Background Technology

[0002] In the field of digital image rendering, the rendering effect of transparent material overlays is a key factor affecting the visual realism of scenes such as games and VR / AR. Traditional rasterization rendering uses a depth-sorting-alpha blending strategy to render transparent objects: first, the transparent objects are sorted from farthest to near according to their depth value from the camera, and then they are drawn and their colors are blended sequentially. However, this approach has significant drawbacks in scenes with multiple transparent objects overlaying: In dynamic scenes, the position and shape of transparent objects change frame by frame, requiring repeated depth calculations and sorting. When the number of objects increases or there is complex occlusion, the computational power consumption for sorting increases exponentially, directly leading to a decrease in the rendering frame rate and making it difficult to meet the needs of real-time interaction.

[0003] Single Z-axis depth sorting cannot characterize the spatial and optical differences of overlapping and nested transparent objects, easily leading to visual artifacts such as hierarchical disorder and through-display. Without establishing a mapping between spatial blocks and objects, full detection is required to screen effective objects within the view frustum, resulting in a high proportion of invalid computations and further exacerbating the computational burden.

[0004] Existing technologies have shortcomings in classification accuracy, depth calculation efficiency, overlap sorting accuracy, and the targeting of view frustum selection. There is an urgent need for a solution that achieves efficient and accurate rendering through scene partitioning, accurate classification, depth optimization, and overlap correction.

[0005] For intersecting or nested transparent objects, a single depth value cannot achieve accurate global sorting, leading to rendering issues such as layer confusion and transparent display, which disrupts the visual realism of virtual scenes and may even cause motion sickness in VR / AR scenarios. In summary, existing rasterization rendering technology struggles to balance rendering efficiency and quality for scenes with multiple transparent objects overlapping, and urgently needs optimization and improvement.

[0006] Therefore, the present invention provides a method, apparatus and system for rendering digital content of images. Summary of the Invention

[0007] In order to overcome the shortcomings of the prior art, at least one technical problem raised in the background art is solved.

[0008] The technical solution adopted by this invention to solve its technical problem is: One objective of this invention is to provide a method for rendering digital content of an image, comprising: Get all transparent objects in the rendering scene, classify and label the transparent objects to form a classification label set, divide the rendering scene into several equal rendering space blocks, and equip each equal rendering space block with a unique classification label set. Obtain the depth value range of static transparent objects in the rendering scene and form a depth value table. Based on the depth value corresponding to the predicted value of the motion trajectory of dynamic transparent objects, establish a dynamic depth value prediction model. Select equal rendering space blocks within the camera's field of view, and sort the rendering order of transparent objects to form a rendering sorting table. In the rendering sorting table, combinations of transparent objects that overlap are selected to form overlapping transparent object combinations. The sorting and correction of these overlapping transparent object combinations are then performed to determine the rendering order of the transparent objects.

[0009] As a further improvement of the present invention, the specific process of forming the classification label set is as follows: Obtain objects in the rendering scene with an alpha value < 1, obtain transparent objects, and assign a unique ID to each transparent object. If a transparent object is bound to a motion controller or has vertex animation, it is marked as a dynamic transparent object. If a transparent object is not bound to a motion controller or has vertex animation, it is marked as a static transparent object. Calculate the predicted motion trajectory value of the transparent object based on the motion controller parameters or animation keyframes. If the geometric vertex coordinates of the transparent object are fixed with the predicted motion trajectory value, it is a rigid transparent object. If the geometric vertex coordinates of the transparent object are not fixed with the predicted motion trajectory value, it is a flexible transparent object. Transparent objects with volume are classified as volume objects, and transparent objects with two-dimensional transparent surfaces are classified as surface objects. They are then integrated according to their ID labels: dynamic transparent objects / static transparent objects / rigid transparent objects / flexible transparent objects / volume objects / surface objects, to obtain a set of classification labels.

[0010] As a further improvement of the present invention, the specific process of equipping each uniform rendering space block with a dedicated set of classification tags is as follows: An octree space partitioning algorithm is used to divide the rendering scene into several equal rendering space blocks. All transparent objects in any equal rendering space block are obtained, and a unique classification label set is established for the equal rendering space block according to the transparent object's ID number - dynamic transparent object / static transparent object - rigid transparent object / flexible transparent object - volume class / plane class.

[0011] As a further improvement of the present invention, the specific process of forming the depth value table is as follows: Extract the ID number of the static transparent object. Using the world coordinate system of the current rendering scene as the reference, and the Z-axis depth of the camera view as the basis for depth calculation, perform coordinate transformation from the world coordinate system to the camera coordinate system for each geometric vertex coordinate, calculate the Z-axis depth value of each geometric vertex coordinate, and filter out the minimum and maximum values ​​from all geometric vertex coordinate depth values ​​to form the depth value range of the static object.

[0012] As a further improvement of the present invention, the specific process of establishing the dynamic depth value prediction model is as follows: Obtain the motion controller parameters, animation keyframes, and physics engine constraints corresponding to the dynamic transparent object; set a time window; and extract the world coordinate system position corresponding to all animation keyframes within the time window for the dynamic transparent object with bound animation keyframes. For dynamic transparent objects with physics engine constraints, calculate the predicted motion trajectory of the dynamic transparent object within a time window; for dynamic transparent objects controlled by user interaction, predict the predicted motion trajectory of the dynamic transparent object within a time window. For all dynamic transparent objects within the time window, the trajectory points in the predicted motion trajectory values ​​are transformed from the world coordinate system to the camera coordinate system. The Z-axis depth value of each trajectory point is calculated. From the depth values ​​of all trajectory points, the minimum and maximum values ​​are selected to form the motion trajectory depth range of the dynamic transparent objects.

[0013] As a further improvement of the present invention, the specific process of forming the rendering sorting table is as follows: The camera parameters of the current camera are obtained, including: field of view, near clipping plane distance, far clipping plane distance, camera position and orientation. Six clipping planes of the view frustum are constructed. An axis-aligned bounding box (AABB) and view frustum intersection test algorithm is used to determine whether the bounding box of the equal rendering space block intersects with the six clipping planes of the view frustum. The equal rendering space blocks that intersect are filtered out. The maximum depth value of the static transparent object is obtained and sorted in descending order of maximum depth value to generate a sorted subqueue of the maximum depth value of the static object. Obtain the real-time depth value and motion trajectory prediction value of the dynamic transparent object, calculate the depth stability weight = deviation between the real-time depth value and the motion trajectory prediction value / preset deviation threshold, insert the maximum depth value corresponding to the maximum depth value sorting subqueue of static objects into the dynamic transparent object maximum depth value sorting subqueue, and mark the dynamic transparent object with a sudden depth change as high priority and insert it at the front of the static object maximum depth value sorting subqueue to form a rendering sorting table.

[0014] As a further improvement of the present invention, the specific process of forming the overlapping transparent object combination is as follows: Obtain the ID numbers of all transparent objects in the rendering sort table, extract the coordinate parameters of the OBB bounding boxes of the transparent objects, and use the Separating Axis Theorem (SAT) to filter out pairs of transparent objects that overlap in x, image processing, and z-axis projections. These pairs are marked as cross candidate groups. For the transparent objects in the cross candidate groups, obtain the pixel points on the surface of the transparent objects, verify the spatial relationship between the transparent objects using ray casting, and mark the intersection range between the transparent objects to form a combination of overlapping transparent objects.

[0015] As a further improvement of the present invention, the specific process of determining the rendering order of transparent objects is as follows: In the combination of overlapping transparent objects, the actual depth value from the pixel of the overlapping transparent object to the camera imaging plane, the optical thickness corresponding to the pixel of the overlapping transparent object, and the contribution of the pixel of the overlapping transparent object are calculated by the ray casting algorithm. The score value is calculated as follows: score value = actual depth value × 0.5 + optical thickness × 0.3 + contribution × 0.2. The transparent objects are then sorted in descending order of score value to obtain the rendering order.

[0016] The second objective of this invention is to provide an image digital content rendering system, comprising: Space Block Mapping Module: Obtain all transparent objects in the rendering scene, classify and label the transparent objects to form a classification label set, divide the rendering scene into several equal rendering space blocks, and equip each equal rendering space block with a unique classification label set. Transparent object depth data processing module: Obtain the depth value range of static transparent objects in the rendering scene and form a depth value table; establish a dynamic depth value prediction model based on the depth value corresponding to the predicted value of the motion trajectory of dynamic transparent objects. View frustum space filtering and initial sorting module: Filters out equal rendering space blocks within the camera's field of view, and sorts the rendering order of transparent objects to form a rendering sorting table; Overlapping Object Combination Correction Module: In the rendering sorting table, the module filters out overlapping transparent object combinations to form overlapping transparent object combinations, sorts and corrects these combinations, and determines the rendering order of the transparent objects.

[0017] The third aspect of this invention is that it provides an image digital content rendering apparatus, comprising: a data acquisition unit, a spatial partitioning and indexing unit, a depth calculation and prediction unit, and a sorting and correction unit.

[0018] The beneficial effects of this invention are as follows: 1. An octree spatial partitioning and multi-dimensional classification labeling of transparent objects are adopted. Each spatial block is equipped with a dedicated classification set. Only the valid spatial blocks and their objects within the view frustum are calculated, avoiding global traversal and repeated processing of invalid objects. The depth value table of static transparent objects supports inter-frame reuse. The depth range of dynamic objects is predicted in advance by the model, reducing the full depth calculation frame by frame, greatly reducing computing power consumption, and ensuring stable rendering frame rate in scenes with multiple transparent objects superimposed.

[0019] 2. Solving the problem of overlapping and nested sorting, improving visual realism. Abandoning a single depth-based sorting logic, it accurately filters overlapping object combinations using directional bounding boxes (OBB) and the separating axis theorem (SAT), and then achieves pixel-level sorting correction based on a weighted scoring method. This completely resolves the hierarchical chaos and transparency issues of intersecting and nested transparent objects, avoiding user dizziness in VR / AR scenarios and restoring the realistic optical effects of transparent materials. Attached Figure Description

[0020] The invention will now be further described with reference to the accompanying drawings.

[0021] Figure 1 This is a flowchart of the steps of an image digital content rendering method according to the present invention; Figure 2 This is a system module diagram of an image digital content rendering system according to the present invention; Figure 3 This is a device connection diagram of an image digital content rendering apparatus according to the present invention. Detailed Implementation

[0022] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0023] Example 1 like Figure 1 As shown in the embodiment of the present invention, an image digital content rendering method includes: Step S10: Obtain all transparent objects in the rendering scene, classify and label the transparent objects to form a classification label set, divide the rendering scene into several equal rendering space blocks, and equip each equal rendering space block with a unique classification label set. In step S10, the specific process of obtaining all transparent objects in the rendering scene is as follows: In the resource management module of the scene rendering engine, all transparent objects (i.e., objects with an alpha value < 1) in the rendering scene are obtained, and a unique ID label is assigned to each of the obtained transparent objects. In step S10, the specific process of classifying and labeling transparent objects to form a classification label set is as follows: Classify each transparent object and detect whether the transparent object is bound to a motion controller or has vertex animation; If a transparent object is bound to a motion controller or has vertex animation, it is marked as a dynamic transparent object; if a transparent object is not bound to a motion controller or has no vertex animation, it is marked as a static transparent object. Calculate the predicted motion trajectory of the transparent object based on motion controller parameters or animation keyframes; Based on the predicted motion trajectory, determine whether the geometric vertex coordinates of the dynamic transparent object are fixed with the predicted motion trajectory. If the geometric vertex coordinates of the transparent object are fixed with the predicted motion trajectory, then the transparent object is marked as a rigid transparent object (such as a glass ball). If the geometric vertex coordinates of the transparent object are not fixed with the predicted motion trajectory (such as transparent fabric), then the transparent object is marked as a flexible transparent object. Transparent objects are classified according to their three-dimensional volume properties. Transparent objects with volume (such as glass containers and bubbles) are marked as volume objects, while transparent objects that are only two-dimensional transparent sheets (such as transparent UIs and gauze) are marked as sheet objects. Based on the ID labels of transparent objects—dynamic transparent objects / static transparent objects—rigid transparent objects / flexible transparent objects—volume class / surface class—we can integrate them to obtain a set of classification labels. In step S10, the specific process of dividing the rendering scene into several equal rendering space blocks and equipping each equal rendering space block with a unique set of classification tags is as follows: The octree space partitioning algorithm is used to divide the rendering scene into several equal rendering space blocks. All transparent objects in any equal rendering space block are obtained. Based on the ID number of the transparent objects, a unique classification label set is established for the equal rendering space block according to the ID number of the transparent objects - dynamic transparent objects / static transparent objects - rigid transparent objects / flexible transparent objects - volume class / pane class. It should be noted that establishing a set of classification tags and equipping each equal rendering space block with a dedicated set of classification tags can reduce the amount of computation when the camera position changes during rendering. Step S20: Obtain the depth value range of static transparent objects in the rendering scene and form a depth value table. Based on the depth value corresponding to the predicted value of the motion trajectory of dynamic transparent objects, establish a dynamic depth value prediction model. In step S20, the specific process of obtaining the depth value range of the static transparent object in the rendering scene and forming a depth value table is as follows: Extract the ID numbers of all objects labeled as static transparent objects from the classification tag set. Using the world coordinate system of the current rendering scene as the reference, uniformly adopt the Z-axis depth of the camera view (i.e., the vertical distance from the object to the camera imaging plane) as the basis for depth calculation to ensure that the calculation dimension of the depth value of all static transparent objects is consistent. Obtain the geometric vertex coordinates of the static transparent object (obtained from step S10), perform a coordinate transformation from the world coordinate system to the camera coordinate system for each geometric vertex coordinate, and calculate the Z-axis depth value of each geometric vertex coordinate; The minimum (Zmin) and maximum (Zmax) values ​​are selected from the depth values ​​of all geometric vertex coordinates. The minimum Zmin and maximum Zmax constitute the depth value range (Zmin-Zmax) of the static object. If the static transparent object is a facet, then directly calculate the depth values ​​of the four vertices of the static transparent object and determine the range, and establish a depth value table. The depth value table includes: static transparent object ID, Zmin value, Zmax value, depth range length, and last update timestamp. In step S20, the specific process of establishing a dynamic depth value prediction model based on the depth value corresponding to the predicted motion trajectory of the dynamic transparent object is as follows: Extract the ID numbers of all objects labeled as dynamically transparent from the classification tag set, and obtain the motion controller parameters, animation keyframes, and physics engine constraints corresponding to the dynamically transparent objects. Based on the interactive requirements of the rendering scene, a time window is set (e.g., 60 frames of animation within 1 second in a game scene; 30 frames within 0.5 seconds in a VR scene) to ensure that the time window covers the movement range of the dynamic transparent object. For the dynamic transparent object with bound animation keyframes, the world coordinate system position corresponding to all animation keyframes within the time window is directly extracted. For dynamic transparent objects (such as floating bubbles) that have physics engine constraints, the physics engine is used to simulate and calculate the predicted motion trajectory of the dynamic transparent objects within a time window. For dynamic transparent objects controlled by user interaction (such as moving shields), the predicted trajectory of the dynamic transparent object within the time window is predicted based on the historical movement speed and direction of the dynamic transparent object. For all dynamic transparent objects within the time window, perform a world coordinate system to camera coordinate system transformation on the predicted motion trajectories of the trajectory points, and calculate the Z-axis depth value of each trajectory point. From the depth values ​​of all trajectory points, the minimum value (Zmin_d image processing namic) and the maximum value (Zmax_d image processing namic) are selected. The minimum value Zmin_d image processing namic and the maximum value Zmax_d image processing namic constitute the motion trajectory depth range of the dynamic transparent object (Zmin_d image processing namic - Zmax_d image processing namic). A dynamic depth prediction model is constructed using a fusion model of temporal LSTM and linear regression. The input features of the dynamic depth prediction model are set as timestamp, current frame depth value, motion speed, motion direction, and current camera position. The output feature is the predicted depth value of the next frame. Based on the trajectory point data within the time window, the training set and validation set are divided in an 8:2 ratio to train the dynamic depth prediction model. After verifying the accuracy of the dynamic depth prediction model (e.g., MSE < 0.01), the dynamic depth prediction model is obtained. Step S30: Select equal rendering space blocks within the camera's field of view, and sort the rendering order of transparent objects to form a rendering sorting table. In step S30, the process of selecting equally spaced rendering blocks within the camera's field of view and sorting the rendering order of transparent objects to form a rendering sorting table is as follows: Obtain any equal rendering space block, extract the world coordinate system bounding box parameters (minimum vertex coordinates, maximum vertex coordinates) of the equal rendering space block, obtain the camera parameters of the current camera, including: field of view, near clipping plane distance, far clipping plane distance, camera position and orientation, construct 6 clipping planes of the view frustum based on the camera parameters (including left, right, top, bottom, near, and far planes), use the axis-aligned bounding box (AABB) and view frustum intersection test algorithm, traverse all equal rendering space blocks, determine whether the bounding box of the equal rendering space block intersects with the 6 clipping planes of the view frustum (including completely inside the view frustum and partially overlapping), filter out equal rendering space blocks that intersect within the view frustum and at the edge, and exclude equal rendering space blocks that are completely outside the view frustum; For each equal rendering space block, obtain the exclusive classification tag set of the equal rendering space block, extract all transparent object ID numbers contained in the exclusive classification tag set, obtain the maximum depth value (Zmax) of static transparent objects from the depth value table, sort them in descending order of maximum depth value (i.e. from far to near), and generate a sorted subqueue of maximum depth value of static objects. The system obtains the real-time depth value and predicted motion trajectory value of the dynamic transparent object within the depth range of its motion trajectory. It calculates the depth stability weight as the deviation between the real-time depth value and the predicted motion trajectory value, divided by a preset deviation threshold. The objects are sorted from high to low according to their depth stability weight. Dynamic transparent objects with stable depth (depth stability weight ≥ 0.7) are inserted into the maximum depth value sorting subqueue corresponding to the maximum depth value of static objects. Dynamic transparent objects with abrupt depth changes (depth stability weight < 0.7) are marked as high priority and uniformly inserted at the front of the maximum depth value sorting subqueue of static objects (sorted from small to large according to depth stability weight), forming a rendering sorting table. Step S40: In the rendering sorting table, filter out the groups of transparent objects that overlap to form overlapping transparent object groups, sort and correct the overlapping transparent object groups, and determine the rendering order of transparent objects. In step S40, the specific process of filtering out overlapping transparent object combinations from the rendering sorting table to form overlapping transparent object combinations is as follows: In the rendering sorting table, obtain the ID number of all transparent objects in the rendering sorting table, extract the bounding box (OBB) coordinate parameters of the transparent objects, and filter out the transparent objects that satisfy the x, image processing, and z axis projection overlap by pairwise combinations and mark them as cross candidate groups. For transparent objects marked as cross candidate groups, obtain the pixels on the surface of the transparent objects (sampled at 10×10 pixel grids), verify the spatial relationship between the transparent objects by ray casting method, mark the cross range between the transparent objects, form a transparent object group, and denoted as the overlapping transparent object group. In step S40, the specific process of sorting and correcting the overlapping transparent object combinations to determine the rendering order of the transparent objects is as follows: In the combination of overlapping transparent objects, the actual depth value from the pixel of the overlapping transparent object to the camera imaging plane is calculated by the ray casting algorithm. The optical thickness corresponding to the pixel of the overlapping transparent object is calculated (optical thickness = thickness of the overlapping transparent object at the pixel position × refractive index; for volumetric objects, it is calculated by the intersection length of the ray and the object; for sheet-like objects, a fixed thickness coefficient is used). The contribution of the pixel of the overlapping transparent object is calculated (contribution = basic transparency of the overlapping transparent object × cosine value of the incident angle of the ray; the incident angle is calculated by the angle between the pixel normal and the ray direction). The weighted summation method is used to calculate the score value. The score value = actual depth value × 0.5 + optical thickness × 0.3 + contribution × 0.2. The transparent objects are rendered in descending order of their scores.

[0024] Example 2 like Figure 2As shown, based on the implementation of Embodiment 1, the present invention provides an image digital content rendering system, comprising: Space Block Mapping Module: Obtain all transparent objects in the rendering scene, classify and label the transparent objects to form a classification label set, divide the rendering scene into several equal rendering space blocks, and equip each equal rendering space block with a unique classification label set. Transparent object depth data processing module: Obtain the depth value range of static transparent objects in the rendering scene and form a depth value table; establish a dynamic depth value prediction model based on the depth value corresponding to the predicted value of the motion trajectory of dynamic transparent objects. View frustum space filtering and initial sorting module: Filters out equal rendering space blocks within the camera's field of view, and sorts the rendering order of transparent objects to form a rendering sorting table; Overlapping Object Combination Correction Module: In the rendering sorting table, the module filters out overlapping transparent object combinations to form overlapping transparent object combinations, sorts and corrects these combinations, and determines the rendering order of the transparent objects.

[0025] Example 3 like Figure 3 As shown, based on the specific implementations of Embodiments 1 and 2, the present invention provides an image digital content rendering apparatus, comprising: Data acquisition unit: Communicates with the resource management module of the scene rendering engine to acquire the geometric information, optical parameters, motion parameters of transparent objects, and the camera's frustum parameters, and outputs a set of classification labels and basic data of transparent objects; Space partitioning and indexing unit: Equipped with an octree space partitioning algorithm chip, it enables fast partitioning of the rendering scene, establishes an index mapping between equal rendering space blocks and transparent objects, and outputs equal rendering space blocks and exclusive classification label sets. Depth Calculation and Prediction Unit: Includes a static depth value calculation subunit and a dynamic depth value prediction subunit. The static depth value calculation subunit calculates the depth value range of static objects based on a coordinate system transformation algorithm. The dynamic depth value prediction subunit integrates a hardware acceleration circuit for a fusion model of LSTM and linear regression, and outputs the predicted depth value of dynamic objects in real time. Sorting and Correction Unit: Equipped with a bounding box overlap detection chip and a pixel-by-pixel scoring calculation unit, it realizes the recognition and accurate sorting of cross regions and outputs the final rendering order of transparent objects.

[0026] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for rendering digital content of an image, characterized in that: include: Get all transparent objects in the rendering scene, classify and label the transparent objects to form a classification label set, divide the rendering scene into several equal rendering space blocks, and equip each equal rendering space block with a unique classification label set. Obtain the depth value range of static transparent objects in the rendering scene and form a depth value table. Based on the depth value corresponding to the predicted value of the motion trajectory of dynamic transparent objects, establish a dynamic depth value prediction model. Select equal rendering space blocks within the camera's field of view, and sort the rendering order of transparent objects to form a rendering sorting table. In the rendering sorting table, combinations of transparent objects that overlap are selected to form overlapping transparent object combinations. The sorting and correction of these overlapping transparent object combinations are then performed to determine the rendering order of the transparent objects.

2. The image digital content rendering method according to claim 1, characterized in that: The specific process for forming the classification label set is as follows: Obtain objects in the rendering scene with an alpha value < 1, obtain transparent objects, and assign a unique ID to each transparent object. If a transparent object is bound to a motion controller or has vertex animation, it is marked as a dynamic transparent object. If a transparent object is not bound to a motion controller or has vertex animation, it is marked as a static transparent object. Calculate the predicted motion trajectory value of the transparent object based on the motion controller parameters or animation keyframes. If the geometric vertex coordinates of the transparent object are fixed with the predicted motion trajectory value, it is a rigid transparent object. If the geometric vertex coordinates of the transparent object are not fixed with the predicted motion trajectory value, it is a flexible transparent object. Transparent objects with volume are classified as volume objects, and transparent objects with two-dimensional transparent surfaces are classified as surface objects. They are then integrated according to their ID labels: dynamic transparent objects / static transparent objects / rigid transparent objects / flexible transparent objects / volume objects / surface objects, to obtain a set of classification labels.

3. The image digital content rendering method according to claim 1, characterized in that: The specific process of equipping each uniform rendering space block with a unique set of classification tags is as follows: An octree space partitioning algorithm is used to divide the rendering scene into several equal rendering space blocks. All transparent objects in any equal rendering space block are obtained, and a unique classification label set is established for the equal rendering space block according to the transparent object's ID number - dynamic transparent object / static transparent object - rigid transparent object / flexible transparent object - volume class / plane class.

4. The image digital content rendering method according to claim 1, characterized in that: The specific process for forming the depth value table is as follows: Extract the ID number of the static transparent object. Using the world coordinate system of the current rendering scene as the reference, and the Z-axis depth of the camera view as the basis for depth calculation, perform coordinate transformation from the world coordinate system to the camera coordinate system for each geometric vertex coordinate, calculate the Z-axis depth value of each geometric vertex coordinate, and filter out the minimum and maximum values ​​from all geometric vertex coordinate depth values ​​to form the depth value range of the static object.

5. The image digital content rendering method according to claim 1, characterized in that: The specific process for establishing the dynamic depth value prediction model is as follows: Obtain the motion controller parameters, animation keyframes, and physics engine constraints corresponding to the dynamic transparent object; set a time window; and extract the world coordinate system position corresponding to all animation keyframes within the time window for the dynamic transparent object with bound animation keyframes. For dynamic transparent objects with physics engine constraints, calculate the predicted motion trajectory of the dynamic transparent object within a time window; for dynamic transparent objects controlled by user interaction, predict the predicted motion trajectory of the dynamic transparent object within a time window. For all dynamic transparent objects within the time window, the trajectory points in the predicted motion trajectory values ​​are transformed from the world coordinate system to the camera coordinate system. The Z-axis depth value of each trajectory point is calculated. From the depth values ​​of all trajectory points, the minimum and maximum values ​​are selected to form the motion trajectory depth range of the dynamic transparent objects.

6. The image digital content rendering method according to claim 1, characterized in that: The specific process for forming the rendering sorting table is as follows: The camera parameters of the current camera are obtained, including: field of view, near clipping plane distance, far clipping plane distance, camera position and orientation. Six clipping planes of the view frustum are constructed. An axis-aligned bounding box (AABB) and view frustum intersection test algorithm is used to determine whether the bounding box of the equal rendering space block intersects with the six clipping planes of the view frustum. The equal rendering space blocks that intersect are filtered out. The maximum depth value of the static transparent object is obtained and sorted in descending order of maximum depth value to generate a sorted subqueue of the maximum depth value of the static object. Obtain the real-time depth value and motion trajectory prediction value of the dynamic transparent object, calculate the depth stability weight = deviation between the real-time depth value and the motion trajectory prediction value / preset deviation threshold, insert the maximum depth value corresponding to the maximum depth value sorting subqueue of static objects into the dynamic transparent object maximum depth value sorting subqueue, and mark the dynamic transparent object with a sudden depth change as high priority and insert it at the front of the static object maximum depth value sorting subqueue to form a rendering sorting table.

7. The image digital content rendering method according to claim 6, characterized in that: The specific process for forming the combination of overlapping transparent objects is as follows: Obtain the ID numbers of all transparent objects in the rendering sort table, extract the coordinate parameters of the OBB bounding boxes of the transparent objects, and use the Separating Axis Theorem (SAT) to filter out transparent objects that overlap in the projection of the x, y, and z axes. These are then paired and marked as cross candidate groups. For the transparent objects in the cross candidate groups, obtain the pixels on the surface of the transparent objects, verify the spatial relationship between the transparent objects using ray casting, and mark the intersection range between the transparent objects to form a combination of overlapping transparent objects.

8. The image digital content rendering method according to claim 7, characterized in that: The specific process for determining the rendering order of transparent objects is as follows: In the combination of overlapping transparent objects, the actual depth value from the pixel of the overlapping transparent object to the camera imaging plane, the optical thickness corresponding to the pixel of the overlapping transparent object, and the contribution of the pixel of the overlapping transparent object are calculated by the ray casting algorithm. The score value is calculated as follows: score value = actual depth value × 0.5 + optical thickness × 0.3 + contribution × 0.

2. The transparent objects are then sorted in descending order of score value to obtain the rendering order.

9. An image digital content rendering system, used to perform the method according to any one of claims 1-8, characterized in that: include: Space Block Mapping Module: Obtain all transparent objects in the rendering scene, classify and label the transparent objects to form a classification label set, divide the rendering scene into several equal rendering space blocks, and equip each equal rendering space block with a unique classification label set. Transparent object depth data processing module: Obtain the depth value range of static transparent objects in the rendering scene and form a depth value table; establish a dynamic depth value prediction model based on the depth value corresponding to the predicted value of the motion trajectory of dynamic transparent objects. View frustum space filtering and initial sorting module: Filters out equal rendering space blocks within the camera's field of view, and sorts the rendering order of transparent objects to form a rendering sorting table; Overlapping Object Combination Correction Module: In the rendering sorting table, the module filters out overlapping transparent object combinations to form overlapping transparent object combinations, sorts and corrects these combinations, and determines the rendering order of the transparent objects.

10. An image digital content rendering apparatus for executing the method according to any one of claims 1-8 and the system according to claim 9, characterized in that: include: The system comprises a data acquisition unit, a spatial partitioning and indexing unit, a depth calculation and prediction unit, and a sorting and correction unit.

Images